Integrate with WebDriver and WebDriver BiDi for automation by nondebug · Pull Request #224 · w3c/gamepad

nondebug · 2025-06-12T22:11:52Z

Closes #???

The following tasks have been completed:

Modified Web platform tests (link to pull request)

Implementation commitment:

WebKit (https://bugs.webkit.org/show_bug.cgi?id=)
Chromium (https://bugs.chromium.org/p/chromium/issues/detail?id=)
Gecko (https://bugzilla.mozilla.org/show_bug.cgi?id=)

index.html

Update webdriver CDDL type definitions. Move all bounds information into SimulateGamepadParams. Remove vendor/product/name parameters. Remove "flush" action and move its functionality to "pause". Add steps for processing and dispatching actions.

Co-authored-by: Marcos Cáceres <marcos@marcosc.com>

sadym-chromium · 2025-09-22T11:11:56Z

index.html

+        </p>
+        <p>
+          An [=input value=] may have an associated <dfn>HID usage identifier</dfn>, a 32-bit value that identifies the type of data represented by the input value.
+          A HID usage does not precisely describe the [=input control layout=], but by convention many [=gamepads=] with similar layouts use similar usages.


by convention many [=gamepads=] with similar layouts use similar usages

It would be great to link a resource with the list of conventional HID usage identifiers. Even if it is not strictly standardizes, it would be helpful to have such a list for reference.

sadym-chromium · 2025-09-22T11:17:41Z

index.html

+        time. A [=touch point=] has a <dfn>touch x coordinate</dfn>
+        and a <dfn>touch y coordinate</dfn> representing the position in the [=touch surface=]'s coordinate system.
+        If a [=touch surface=] is on the top, bottom, front, or back side of the [=gamepad=] then the [=touch x coordinate=] is measured along the left-right axis, otherwise it is measured along the top-bottom axis.
+        The [=touch y coordinate=] is measured along the perpendicular axis.


NIT: should we define the specific orientation of the the perpendicular axis?

index.html

sadym-chromium · 2025-09-22T11:27:46Z

index.html

+        </p>
+        <ol>
+          <li>
+            Return 2 (|value| − |minimum|) / (|maximum| − |minimum|) − 1.


I would expect the normalized output of input with min 1, max 5 and val 3 to be 1/2:

Suggested change

Return 2 (|value| − |minimum|) / (|maximum| − |minimum|) − 1.

Return (|value| − |minimum|) / (|maximum| − |minimum|).

sadym-chromium · 2025-09-22T11:38:32Z

index.html

          To <dfn>map and normalize axes</dfn> for |gamepad:Gamepad|, run the
          following steps:


NIT: The definition below now can be formulated in strict terms introduced previously: input values #dfn-input-values and axis list #dfn-axis-list

<li>Let |axisValues:list| be a [=list=] of {{unsigned long}} values representing the most recent logical axis input values for each axis input of the device represented by |gamepad|. </li>```

Also mappedIndex is not used, so probably it can be omitted.

sadym-chromium · 2025-09-22T11:47:17Z

index.html

                <li>[=Issue a haptic effect=] to the actuator with |type|,
                |params|, and the |playEffectTimestamp|.
                </li>
+                <li>If |this|'s [=gamepad=] is a [=simulated gamepad=], run the steps to [=vibrate a simulated gamepad=] for |this|'s [=gamepad=] with |params|.
+                </li>


sadym-chromium · 2025-09-22T11:56:37Z

index.html

      </h3>
      <p>
-        When a gamepad becomes available on the system, run the following
+        <dfn>When a gamepad becomes available on the system</dfn>, run the following


I believe it's more correct to split this into a trigger and a process.

Suggested change

<dfn>When a gamepad becomes available on the system</dfn>, run the following

When a [=gamepad=] <var>gamepad</var> becomes available on the system,

[=emit `gamepadconnected` event=] for <var>gamepad</var>.

</p>

<p>

To <dfn>emit `gamepadconnected` event</dfn> for given [=gamepad=] <var>gamepad</var>, run the following

Also I believe the [=current global object=] is ambiguous, especially in case of simulated devices.

sadym-chromium · 2025-09-22T11:58:20Z

index.html

      </h3>
      <p>
-        When a gamepad becomes unavailable on the system, run the following
+        <dfn>When a gamepad becomes unavailable on the system</dfn>, run the following


The same issue as with When a gamepad becomes available on the system.

sadym-chromium · 2025-09-22T12:00:36Z

index.html

+        It is OPTIONAL for a [=user agent=] to support them.
+      </p>
+      <p>
+        Each [=top-level traversable=] has a <dfn>simulated gamepad map</dfn> which is an [=ordered map=] of [=version 4 UUIDs=] to [=gamepad input sources=].


I think the UUID key it sufficient and we don't need to have it ordered map.

Suggested change

Each [=top-level traversable=] has a <dfn>simulated gamepad map</dfn> which is an [=ordered map=] of [=version 4 UUIDs=] to [=gamepad input sources=].

Each [=top-level traversable=] has a <dfn>simulated gamepad map</dfn> which is a [=map=] of [=version 4 UUID=] and [=gamepad input source=].

sadym-chromium · 2025-09-22T12:04:22Z

index.html

+        It is OPTIONAL for a [=user agent=] to support them.
+      </p>
+      <p>
+        Each [=top-level traversable=] has a <dfn>simulated gamepad map</dfn> which is an [=ordered map=] of [=version 4 UUIDs=] to [=gamepad input sources=].


Perhaps it's simulated gamepad, not gamepad input source:

Suggested change

Each [=top-level traversable=] has a <dfn>simulated gamepad map</dfn> which is an [=ordered map=] of [=version 4 UUIDs=] to [=gamepad input sources=].

Each [=top-level traversable=] has a <dfn>simulated gamepad map</dfn> which is an [=ordered map=] of [=version 4 UUIDs=] to [=simulated gamepad=].

I believe the UUID version is important only when it is generated:

Suggested change

Each [=top-level traversable=] has a <dfn>simulated gamepad map</dfn> which is an [=ordered map=] of [=version 4 UUIDs=] to [=gamepad input sources=].

Each [=top-level traversable=] has a <dfn>simulated gamepad map</dfn> which is an [=ordered map=] of [=UUID=] to [=gamepad input sources=].

sadym-chromium · 2025-09-22T12:13:05Z

index.html

+            <pre class="cddl">
+              emulation.GamepadBounds = {
+                max: float64,
+                ? min: float64,


You can define .default in CDDL like this:

Suggested change

? min: float64,

? min: float64 .default 0.0,

sadym-chromium · 2025-09-22T12:16:59Z

index.html

+            <pre class="cddl">
+              emulation.SimulateGamepadParams = {
+                type: "gamepad",
+                context: browsingContext.BrowsingContext,


IDK if we can cross-reference CDDL to BiDi spec, so let's use text the same way Bluetooth spec does.

Suggested change

context: browsingContext.BrowsingContext,

context: text,

sadym-chromium · 2025-09-22T12:28:00Z

index.html

+                Let |simulatedGamepads| be the [=simulated gamepad map=].
+              </li>
+              <li>
+                If |simulatedGamepads|[|params|.|deviceId|] does not [=map/exist=], abort these steps.


sadym-chromium · 2025-09-22T12:31:02Z

index.html

+                If |simulatedGamepads|[|params|.|deviceId|] does not [=map/exist=], abort these steps.
+              </li>
+              <li>
+                Let |gamepad| be |simulatedGamepads|[|params|.|deviceId|].


sadym-chromium · 2025-09-22T12:31:14Z

index.html

+                Run the steps [=when a gamepad becomes unavailable on the system=] with |gamepad|.
+              </li>
+              <li>
+                [=map/Remove=] |simulatedGamepads|[|params|.|deviceId|].


sadym-chromium · 2025-09-22T12:36:46Z

index.html

+              }
+            </pre>
+            <p>
+              To get <dfn>the identifier for a simulated gamepad</dfn> |gamepad|:


Optionally, we can also extend the definition of "simulated gamepad" as "Simulated gamepad has an associated [=UUID=] simulated gamepad identifier".

sadym-chromium · 2025-09-22T12:38:28Z

index.html

+                Let |deviceId:string| be [=the identifier for a simulated gamepad=] |gamepad|.
+              </li>
+              <li>
+                If |deviceId| is `null`, abort these steps.


Maybe we want to notify the user that the given gamepad is removed by throwing an exception?

sadym-chromium · 2025-09-22T12:39:31Z

index.html

+                Let |channels| be the [=list=] « `"strongMagnitude"`, `"weakMagnitude"`, `"leftTrigger"`, `"rightTrigger"` ».
+              </li>
+              <li>
+                [=list/For each=] |channel| of |channels|:


It means that we ignore all other potential values. IDK if this is correct, and how extensible it is.

sadym-chromium · 2025-09-22T12:42:30Z

index.html

+                Let |body| be a [=map=] matching the `emulation.SimulatedDeviceEvent` production with the `params` field set to |eventParams|.
+              </li>
+              <li>
+                Let |relatedNavigables| be a [=set=] containing |navigable|.


Note: this means that even if the event was initiated from an iframe, the event will be emitted only on the top-level navigable. I believe this is fine.

sadym-chromium · 2025-09-22T12:44:01Z

index.html

+            Gamepad Input Source
+          </h4>
+          <p>
+            A <dfn>gamepad input source</dfn> is an [=input source=] that is associated with a [=gamepad=]-like device.


Do we need the gamepad input source instance? Cannot it be a simulated gamepad?

sadym-chromium · 2025-09-22T12:47:45Z

index.html

+            </dl>
+          </section>
+          <p>
+            Extend the dispatch action algorithm table in [=dispatch tick actions=] with the following table:


Please add reference to the WebDriver Classic spec, where dispatch tick actions is defined:

spec: WEBDRIVER; urlPrefix: https://w3c.github.io/webdriver/ text: dispatch tick actions; url: dfn-dispatch-tick-actions

sadym-chromium · 2025-09-22T12:51:59Z

index.html

+              Let |simulated gamepads| be the [=simulated gamepad map=].
+            </li>
+            <li>
+              If |simulated gamepads|[|device id|] does not [=map/exist=], abort these steps.


Return a error with "invalid argument" or a new error like "no such simulated gamepad"

sadym-chromium · 2025-09-22T12:55:26Z

index.html

+          <p>
+            A [=gamepad input source=] has the following items:
+          </p>
+          <table>


Consider defining the gamepad input source (or simulated gamepad) as a [=struct=] with [=struct/item=] (example). That would allow for referencing the items with the links to it's definitions and search for references (example).

sadym-chromium · 2025-09-22T13:00:05Z

index.html

+              Let |action| be an [=action object=] constructed with arguments |id|, `"gamepad"`, and |subtype|.
+            </li>
+            <li>
+              If |subtype| is `"pause"`, let |result| be the result of [=trying=] to [=process a pause action=] with arguments |action item| and |action|, and return |result|.


sadym-chromium · 2025-09-22T13:01:36Z

index.html

+            <li>
+              Return [=success=] with data |action|.
+            </li>


Should not it be an "invalid argument" error, as the actions is unknown?

sadym-chromium · 2025-09-22T13:05:57Z

index.html

+            Actions
+          </h4>
+          <p>
+            To <dfn>process a gamepad action</dfn> given |id| and |action item|:


It looks like this part is not finished yet. Please let me know if I can help with it.

Co-authored-by: Maksim Sadym <69349599+sadym-chromium@users.noreply.github.com>

nondebug added 4 commits March 18, 2025 15:32

Add a model section

701a388

Split model sections

cd32d3a

Add a section on input values and define logical values

86f434d

Integrate with WebDriver and WebDriver-BiDi for automation

4336aaa

marcoscaceres reviewed Jun 12, 2025

View reviewed changes

index.html Outdated Show resolved Hide resolved

Align with Gamepad API testing doc

0d8f530

Update webdriver CDDL type definitions. Move all bounds information into SimulateGamepadParams. Remove vendor/product/name parameters. Remove "flush" action and move its functionality to "pause". Add steps for processing and dispatching actions.

nondebug mentioned this pull request Aug 6, 2025

Remove tests that rely on undefined testing APIs web-platform-tests/rfcs#172

Open

nondebug added TPAC2025 Topics for discussion at TPAC 2025 and removed TPAC2025 Topics for discussion at TPAC 2025 labels Aug 27, 2025

Update index.html

1e5e770

Co-authored-by: Marcos Cáceres <marcos@marcosc.com>

sadym-chromium reviewed Sep 22, 2025

View reviewed changes

index.html Outdated Show resolved Hide resolved

sadym-chromium reviewed Sep 22, 2025

View reviewed changes

xingri mentioned this pull request Sep 23, 2025

Advancing Gamepad API Test Automation web-platform-tests/interop#1148

Open

Update index.html

02c8a13

Co-authored-by: Maksim Sadym <69349599+sadym-chromium@users.noreply.github.com>

nondebug mentioned this pull request Oct 30, 2025

TPAC 2025 status report #226

Closed

	Return 2 (\|value\| − \|minimum\|) / (\|maximum\| − \|minimum\|) − 1.
	Return (\|value\| − \|minimum\|) / (\|maximum\| − \|minimum\|).

		To <dfn>map and normalize axes</dfn> for \|gamepad:Gamepad\|, run the
		following steps:

-        <dfn>When a gamepad becomes available on the system</dfn>, run the following
+        When a [=gamepad=] <var>gamepad</var> becomes available on the system,
+        [=emit `gamepadconnected` event=] for <var>gamepad</var>.
+      </p>
+      <p>
+        To <dfn>emit `gamepadconnected` event</dfn> for given [=gamepad=] <var>gamepad</var>, run the following

	Each [=top-level traversable=] has a <dfn>simulated gamepad map</dfn> which is an [=ordered map=] of [=version 4 UUIDs=] to [=gamepad input sources=].
	Each [=top-level traversable=] has a <dfn>simulated gamepad map</dfn> which is a [=map=] of [=version 4 UUID=] and [=gamepad input source=].

	If \|simulatedGamepads\|[\|params\|.\|deviceId\|] does not [=map/exist=], abort these steps.
	If \|simulatedGamepads\|[\|params\|["<code>deviceId</code>"]] does not [=map/exist=], return [=error=]
	with [=error code=] [=invalid argument=].

	Let \|gamepad\| be \|simulatedGamepads\|[\|params\|.\|deviceId\|].
	Let \|gamepad\| be \|simulatedGamepads\|[\|params\|["<code>deviceId</code>"]].

	[=map/Remove=] \|simulatedGamepads\|[\|params\|.\|deviceId\|].
	[=map/Remove=] \|simulatedGamepads\|[\|params\|["<code>deviceId</code>"]].

	If \|subtype\| is `"pause"`, let \|result\| be the result of [=trying=] to [=process a pause action=] with arguments \|action item\| and \|action\|, and return \|result\|.
	If \|subtype\| is `"pause"`, return the result of [=trying=] to [=process a pause action=] \|action item\| and \|action\|.

Conversation

nondebug commented Jun 12, 2025 • edited by pr-preview bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

sadym-chromium Sep 22, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

sadym-chromium Sep 22, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

sadym-chromium Sep 22, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

sadym-chromium Sep 22, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

sadym-chromium Sep 22, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

sadym-chromium Sep 22, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

sadym-chromium Sep 22, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

sadym-chromium Sep 22, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

nondebug commented Jun 12, 2025 •

edited by pr-preview bot

Loading

sadym-chromium Sep 22, 2025 •

edited

Loading

sadym-chromium Sep 22, 2025 •

edited

Loading

sadym-chromium Sep 22, 2025 •

edited

Loading

sadym-chromium Sep 22, 2025 •

edited

Loading

sadym-chromium Sep 22, 2025 •

edited

Loading

sadym-chromium Sep 22, 2025 •

edited

Loading

sadym-chromium Sep 22, 2025 •

edited

Loading

sadym-chromium Sep 22, 2025 •

edited

Loading